The magnetoencephalogram (MEG) is a useful new tool for estimating the location of electrical sources in the brain. To date, these estimates have been made by the feature matching method in which certain spatial features of MEG maps measured on the head are matched to those in a calculated map for a dipole in a model of the head. In this method, the accuracy of the location estimate depends on the accurary with which only a few points in the map can be measured on the head. In addition, the usefulness of this method is the moving dipole in a model of the head. In this method, the accuracy of the location estimate depends on the accurary with which only a few points in the map can be measured on the head. In addition, the usefulness of this method is greatly reduced for locating large, distributed sources. A more accurate method is the moving dipole method in which the total measured map is matched to the calculated map of a moving dipole in a head model; the location orientation, and amplitude of the moving dipole are adjusted until a least squares error fiut between the measured and calculated maps is obtained. This method has not been used with MEG's. The purpose of this research is to investigate the use of the moving dipole method with MEG's for getting information about sources in the brain. This research would determine the accuracy of the moving dipole method using MEG's for locating small, dipolar sources. Computer modeling studies would be performed to determine the effects of noise in the data, measurement location errors, etc. on the accuracy of MEG moving dipole solutions. Comparisons with the effects on EEG moving dipole solutions would be made. This research would also perform computer modeling studies to determine the complementary information about large, distributed sources that MEG moving dipole solutions can provide when compared with EEG moving dipole solutions. The results of the computer modeling studies would be used to analyze moving dipole solutions for three evoked sources in the brain. These sources are the N20 somatosensory source for stimulation of the median nerve, the P100 visual source for stimulation by checkerboard pattern reversals, and the N100-P200 auditory sources for stimulation by a sound click. This analysis would help to clarify the location and nature of these sources